A conventional secondary station (or a mobile station or user equipment (UE)) that is configured to receive HSDPA (High Speed Downlink Packet Access) data packets transmits Channel Quality Indicator (CQI) feedback to the primary station (or a base station, or a NodeB) on a periodic cycle whose period is configured by the network.
In CPC mode (Continuous Packet Connectivity), the control channel overhead is reduced between data packets, so that control signalling is only sent in periodic bursts. This is known as Discontinuous Transmission (or DTX), wherein the control signalling is transmitted intermittently. The reception may also be discontinuous, known as Discontinuous Reception (or DRX).
A communication system comprising a secondary station and a primary station may operate with multiple antennas in a Multiple-Input Multiple-Output (MIMO) mode.
In both MIMO and non-MIMO modes, it is advantageous for the network to align the channel quality indicator feedback cycle period k with the DTX and DRX cycles of the CPC mode.
If this alignment comprises setting k equal to the DTX cycle period, then the CQI feedback rate is unaffected by the value of CQI_DTX_Priority provided that the transmission bursts of the DTX cycle align with the transmission times for CQI in the CQI feedback cycle. CQI_DTX_Priority is a parameter which controls whether the “off” period of the DTX cycle or the transmission of the CQI value takes precedence. For instance, if CQI_DTX_Priority=0, a CQI report will not be sent if the DTX pattern means that the control signalling should be switched off at that time. On the other hand, if CQI_DTX_Priority=1, a CQI report will be transmitted even if the DTX pattern indicates an “off” period.
More usefully, the DTX/DRX cycle period may be configured to be an integer multiple of the CQI feedback cycle period k. In this way the CQI feedback rate is reduced in the intervals between packet transmission, assuming an appropriate setting of the timer CQI_DTX_TIMER which controls the length of time after each packet for which CQI_DTX_Priority=1. This enables an increase in UE sleep time and reduction in interference. Thus, power is saved and interference is reduced. This is for instance illustrated by FIG. 1, which represents time graphs T11, T12 and T13, for an example having CQI feedback cycle period k=2 subframes with DTX period L=4 subframes, and when the CQI_DTX_Priority is 0. T11 represents the DTX pattern, which is high during a transmission phase, i.e. when data may be transmitted, and low otherwise. T12 represents the CQI reports scheduled in a normal use, i.e. not in CPC mode. Normally, a CQI report should be transmitted every second subframe. Finally, T13 represents the CQI reports actually transmitted, i.e. the CQI reports which coincide with a transmission phase of the DTX.
If the DTX cycle period L is longer than k and not an integer multiple of it, a more significant reduction in the CQI reporting rate will occur.
When the UE is not in MIMO mode, the skipping of certain CQI reports causes no particular problems (other than reduced availability of CQI) when the DTX cycle period is greater than k.
However, when the UE is in MIMO mode, multiple different types of CQI reports can be sent—for example two different types designated Type A and Type B, indicating the PCI/CQI (where PCI denotes Precoding Control Information, indicating a preferred beamforming weight for application at the transmitter) for the secondary station's preferred number of streams and the PCI/CQI for a single stream respectively. Therefore skipping some CQI reports due to the DTX cycle potentially has more impact and needs to be considered more carefully. Indeed, when the CPC is not configured, a typical CQI reporting pattern may comprise sending N TypeA reports out of every M reports (the remainder being Type B), where N/M can for example take the following values: {1/2, 2/3, 3/4, 4/5, 5/6, 6/7, 7/8, 8/9, 9/10, 1/1}.
This means that when the DTX cycle period is longer than k, it is highly likely that all the CQI reports of one type will be skipped if the DTX cycle period is an integer multiple of k as in non-MIMO mode.
For example, if N/M=1/2 as depicted on FIG. 2, then setting the DTX cycle period L to any multiple of k will result in either all the Type A reports being skipped or all the Type B reports being skipped (which type is skipped depends on the time offset between the DTX cycle and the cycle of the CQI feedback pattern). On time graph T23, only type A are transmitted and all the type B CQI reports from time graph T22, representing the schedule of transmission of the CQI reports, are skipped.
In general, setting the DTX cycle to any integer multiple of Mk will result in all the CQI reports of one type being skipped. For example, if N/M=3/4, setting the DTX cycle period to 4 k will cause all the transmitted CQI reports to be of the same type.
Reconfiguring N/M every time CQI_DTX_Priority changed would be impractical and is contrary to the spirit of CPC which avoids slow higher-layer signalling for transitions between normal and DTX modes.
Another possibility would be to configure two sets of N/M values when MIMO mode is configured for the secondary station. However, this is problematic as the second set of N/M is only needed when both CPC and MIMO are configured. The control signalling for these two features is independent, so there is no obvious control signalling message in which to include the second set of N/M values.
A solution is to define a Physical-Layer rule which prevents all of one type of CQI report being extinguished by the DTX cycle. One novel rule would be to make the type of the CQI reports when DTX is active depend on the actually-transmitted reports rather than the Connection Frame Number (CFN).
However, this would lose all relation to the Connection Frame Number. Indeed, the existing definition of which of Type A and Type B should be sent is derived directly from the Connection Frame Number which is known to both secondary and primary stations, so that the primary station knows exactly what type of CQI to expect in each frame. By losing all relation to the CFN, this would increase the probability of the primary station “losing synchronisation” with which type of CQI report the secondary station is sending, thus possibly increasing the error rate for the decoding of the CQI reports received at the primary station.